Statements (112)
| Predicate | Object |
|---|---|
| gptkbp:instanceOf |
large language model
|
| gptkbp:architecture |
gptkb:transformation
|
| gptkbp:competitor |
gptkb:OpenAI_GPT
gptkb:Anthropic_Claude gptkb:Google_Gemini gptkb:Meta_Llama |
| gptkbp:context |
32k tokens
65k tokens |
| gptkbp:developedBy |
gptkb:Mistral_AI
|
| gptkbp:firstReleased |
2023
|
| https://www.w3.org/2000/01/rdf-schema#label |
Mistral language models
|
| gptkbp:language |
gptkb:French
English |
| gptkbp:license |
Apache 2.0
|
| gptkbp:mixtureOfExperts |
yes
|
| gptkbp:notableFeature |
scalable architecture
supports multiple languages high performance on benchmarks open weights supports code generation supports function calling supports fine-tuning supports instruction following efficient inference supports code completion supports question answering supports reasoning tasks supports retrieval-augmented generation supports summarization supports tool use optimized for deployment supports ALiBi positional encoding supports API integration supports RLHF supports chat format supports cloud deployment supports code analysis supports code chat supports code debugging supports code documentation supports code explanation supports code formatting supports code generation for C# supports code generation for C++ supports code generation for Go supports code generation for Java supports code generation for JavaScript supports code generation for Kotlin supports code generation for PHP supports code generation for Python supports code generation for Ruby supports code generation for Rust supports code generation for SQL supports code generation for Scala supports code generation for Shell supports code generation for Swift supports code generation for TypeScript supports code generation for multiple languages supports code linting supports code optimization supports code refactoring supports code review supports code search supports code summarization supports code tasks supports code testing supports code translation supports custom datasets supports distributed training supports document understanding supports edge deployment supports efficient memory usage supports evaluation on standard benchmarks supports flash attention supports grouped-query attention supports inference acceleration supports instruction tuning supports large batch training supports low-rank adaptation (LoRA) supports model parallelism supports multi-GPU training supports multi-query attention supports multilingual tasks supports on-premise deployment supports pipeline parallelism supports plugins supports prompt engineering supports prompt tuning supports quantization supports rotary positional embeddings supports safety alignment supports sliding window attention supports streaming supports tensor parallelism |
| gptkbp:notableModel |
gptkb:Mistral-7B
gptkb:Mixtral-8x7B gptkb:Mistral_Large Mistral-8x22B Mistral-8x7B Mixtral-8x22B |
| gptkbp:openSource |
yes
|
| gptkbp:parameter |
7B
46.7B 12B 22B |
| gptkbp:uses |
chatbot
translator code generation summarization text generation |
| gptkbp:bfsParent |
gptkb:Mistral_API
|
| gptkbp:bfsLayer |
7
|